On average, only 1 of 10 breast biopsies identifies a malignancy: the other nine are diagnosed as various types of "benign" breast disease. Yet, individual women with "benign" lesions have a 10-20 fold increased risk of developing breast cancer. There is an obvious need for a predictor or prospective determinant of risk in patients who have histopathologically "benign" breast biopsies. Using a simple assay based on analysis of genetic instability among anonymous microsatellite sequences with pathologically benign lesions, we have identified a number of "benign" lesions as being clonal, genetically unstable, and pre- malignant. In Phase I, we will begin studies to verify that the molecular changes we observe in tissue specimens are prognostically valid. The molecular test that we have developed should have broad clinical and commercial application because of the following features: 1) The test does not depend on the knowledge or identification of specific oncogenes or tumor suppressor genes; 2) The test is easy to perform, requiring only an elementary experience with PCR (already in wide use in clinical laboratories; 3) The test can be performed on archival, paraffin-fixed specimens, as well as on fresher tissue; 4) The test can eventually be applied to tumor specimens other than breast tumors. PROPOSED COMMERCIAL APPLICATION: Histologic analysis of breast lesions using light microscopy has significant limitations. We have established a simple molecular assay which can identify which of a number of "benign" lesions are actually pre-malignant. This molecular test should have broad clinical and commercial application in diagnosis and prognosis of breast lesions. The test can be performed on archival, paraffin-fixed specimens, as well as on fresher tissue.